News / Vancouver

Bus tweets help transit

'The tweets on the bus go round and round…'

Saeid Allahdadian, an engineer and postdoctoral researcher at UBC who has developed a computer tool to collect location data from tweets on transit, stands at the university's transit loop

Coutesy UBC/Paul Joseph

Saeid Allahdadian, an engineer and postdoctoral researcher at UBC who has developed a computer tool to collect location data from tweets on transit, stands at the university's transit loop

Next time you're tempted to roll your eyes at a bus full of passengers whose eyes seem glued to their phones, consider that they could in fact be improving our entire transportation system.

You read that right. A new study from the University of British Columbia has gathered up three million tweets — public messages posted to the social media site Twitter — over three weeks and discovered they hold enough location data to study how people get around our region and when.

"We could recognize and predict an individual's movement patterns based on the tweets they make during a day," explained UBC post-doctoral computer scientist Saeid Allahdadian by phone. "We can predict that at this time of day or this day of the year, this many people are expected to go from this point to this point.

"That's very important for urban planning and other applications, it could help optimize the transportation system to serve them better, but it could also help emergency responders and urban planners prioritize where the greatest number of people might need help."

Of course, the idea of a Minority Report-style prediction model that could tell authorities where anyone would likely be at any given time might give many privacy-minded citizens pause, he admitted.

But because Twitter's privacy settings allow users to choose whether or not to reveal each tweet's location using GPS — known as "geotagging" — it's "completely voluntary," he said. "People choose if they want to share their location info."

And even though just one per cent of tweets in our region are geotagged, they nonetheless offer a massive trove of data that reveals big-picture patterns.

"They can give us an idea of where transit services might need room for improvements, and where they're sufficient," he said. Additionally, what riders tweet about may be no surprise.

"A lot of tweets were directed at TransLink, often about problems," he noted. "But from the models and methods I developed, in realtime TransLink could quickly locate that problem."

Just three weeks that he surveyed this summer, he saw some routes were well-covered — for instance, commuters between Surrey and Vancouver — other highly popular routes saw high tweet traffic but few buses, such as Langley to Surrey or the Broadway corridor to UBC.

More on Metronews.ca